一个图像数据库检索系统的结构设计和快速检索方法

Architecture Design and Fast Retrieving Method of An Image Database Retrieving System

本文介绍了图像数据库区别于常规数据库的特点,并以数字化图书馆的图像数据库子系统Basestar为例,介绍了图像数据库检索系统的结构和检索方法。Basestar采用综合了颜色、形状、纹理三种特征矢量的综合性检索方法,本文详细介绍了这种方法,并论述了在此方法基础上,根据三角形不等式准则改进而成的快速检索方法的原理和实现。实验证明Basestar系统具有较好的检索效果,检索结果证明快速检索方法是有效的。

In this paper, the speciality of an image database different from a general database is described. With the illustration of Basestar, the image database subsytem in digital library, the architecture and retrieving method of the image database retrieving system is stated. Basestar employs retrieving method synthesized with color feature vector, shape feature vector and texture feature vector. In this paper,this synthesized method is described in details,and the fast algorithm on the basis of this method, according to triangle inequality criterion, is also represented. Experiments shows that Basestar has a good retrieving effect and the effectivity of the fast retrieving method is proved by retrieving results.